Markdownee Apify Actor

Use the Apify Actor for hosted URL extraction, configurable link-following crawls, and scheduled runs. Its dataset indexes the outcomes; content can be placed inline or in the key-value store for later API access.

Start a run

Open the Actor page, supply starting URLs, choose settings, and select Start. The command-line equivalent is:

apify call glueo/markdownee \
  --input='{"startUrls": [{"url": "https://example.com"}]}'

The Apify API accepts the same input:

curl -X POST "https://api.apify.com/v2/acts/glueo~markdownee/runs" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"startUrls": [{"url": "https://example.com"}]}'

These commands start Actor runs. An API token belongs in your execution environment, not in shared examples.

Read the input reference

The following tables retain the field names, types, and defaults used by the input contract. A blank default means no fixed schema default; the field's description explains omission behavior.

Crawl settings

Crawl filtering

Content extraction

Output destinations

Browser settings

Proxy

Cookies & headers

Understand the output record

Successful pages use status: 'success'. Exhausted request retries produce failed records with errors, retryCount, and crawl.scrapedAt; crawl.loadedUrl is included when available. With storeSkippedUrls enabled, skips use status: 'skipped' and skipReason.

A *-kvs route stores content separately and references its key and, when available, public url. A *-dataset route includes content in the record. Content nodes carry an MD5 hash and UTF-8 bytes count.

The example below illustrates record structure; its title, timestamps, sizes, and hashes are sample values:

{
  "url": "https://example.com/article",
  "status": "success",
  "metadata": {
    "title": "Article Title",
    "author": "Author Name",
    "date": "2026-01-15",
    "description": "Article description",
    "siteName": "Example",
    "languageCode": "en"
  },
  "crawl": {
    "loadedUrl": "https://example.com/article",
    "scrapedAt": "2026-04-11T12:00:00.000Z",
    "httpStatusCode": 200,
    "depth": 1,
    "referrerUrl": "https://example.com/"
  },
  "original": {
    "hash": "f8e6bd335e04d03e1be6798c2c72349c",
    "bytes": 89898
  },
  "markdown": {
    "hash": "c485356090a92c6a45e8c1155c14d8ee",
    "bytes": 4523,
    "key": "markdown-141fbc787408697a5d22735982be532a.md",
    "url": "https://api.apify.com/v2/key-value-stores/<storeId>/records/<key>"
  }
}

Selected formats appear as txt, markdown, html, or minifiedHtml. The original node carries the crawler capture's hash and size; an original-* route is needed to store its content. This capture precedes content extraction but may already reflect rendering, serialization, or consent handling.

Identify Markdown-sourced content

Enabled markdownDiscovery can accept a representation published by the origin. These records include markdownSource with the source URL and a mechanism of response, alternate, negotiated, or sibling.

verbatim is true when Markdown uses the served body after source front-matter removal and the selected output layout. It is false for Markdown converted from cleaned HTML or when no Markdown format was saved. The record's original is derived from the served representation on this path. Without markdownSource, content came from HTML extraction.

Discovery steps depend on crawler capabilities. Up to three alternates, one negotiated refetch, and one sibling may be attempted, with robots.txt checks and per-origin budgets also affecting requests. A rejected representation falls back to HTML.

Combine filters and destinations

This example sets starting URLs, include/exclude patterns, crawl limits, output routes, and Apify Proxy configuration:

{
  "startUrls": [{"url": "https://example.com/blog"}],
  "globs": [{"glob": "https://example.com/blog/**"}],
  "exclude": [{"glob": "https://example.com/blog/archive/**"}],
  "maxRequestsPerCrawl": 100,
  "maxCrawlDepth": 2,
  "save": ["markdown-kvs", "minified-html-kvs"],
  "mode": "precision",
  "outputLayout": "enhanced",
  "deduplication": "aggressive",
  "proxyConfiguration": {
    "useApifyProxy": true
  }
}

URL patterns filter candidate links; they do not enable link following by themselves. Add a selector such as a[href] when the run should enqueue links from the starting page, or supply the complete URL list.

Related guides

• Help hub lists the interfaces.
• Playground previews a page and prepares settings.
• npm CLI covers local commands and JSON configuration.
• npm library covers Node.js integration.
• PyPI package documents the alpha Python wrapper.

Updated: September 8, 2026